FABS Fish Diversity Summaries
Ben Millard-Martin
Fri Jan 12 14:11:36 2018
Setup
These blocks of code sets base for the subsequent analyses:
- loads all necessary packages
- set working directory
- read in diversity abundance, diversity length, ysi and habitat
- create wide data for Vegan package
- recreate long data with 0 observations
- clean ysi data(remove duplicates)
- create wide temperature data from ysi data
- merge temperature and habitat data with diversity abundance
- assign levels and names to month objects
Load Packages
# load packages
rm(list=ls())
library("tidyverse", lib.loc="~/R/win-library/3.4")
library("vegan", lib.loc="~/R/win-library/3.4")
library("BiodiversityR", lib.loc="~/R/win-library/3.4")
library("knitr", lib.loc="~/R/win-library/3.4")
library("yaml", lib.loc="~/R/win-library/3.4")
library("rmarkdown", lib.loc="~/R/win-library/3.4")
library("magrittr", lib.loc="~/R/win-library/3.4")
library("lubridate", lib.loc="~/R/win-library/3.4")
library("TropFishR", lib.loc="~/R/win-library/3.4")Set working directory
# make fabsdivlength name of and create working file
#setwd("C:/Users/FABS/Desktop/Rwd_Ben/FABS/working_files/data")Read in data
# read in diversity abundnce, diversity length, ysi and habitat data
da <- read.csv ("C:/Users/FABS/Desktop/Rwd_Ben/FABS/working_files/data/da.csv")
dl <- read.csv ("C:/Users/FABS/Desktop/Rwd_Ben/FABS/working_files/data/dl.csv")
habitat <- read.csv ("C:/Users/FABS/Desktop/Rwd_Ben/FABS/working_files/data/habitat.csv")
ysi <- read.csv ("C:/Users/FABS/Desktop/Rwd_Ben/FABS/working_files/data/ysi.csv")Reformat data
- create wide data
- clean YSI data
- extract temperature data
- merge temperature and habitat data with diversity abundance data
- create new month labels
#create da.wide
da.wide <- da %>%
group_by(year, month, day, site, species) %>%
summarise(abundance1 = sum
(abundance, na.rm = TRUE)
)%>%
spread("species", "abundance1")
#replace na with 0
da.wide[is.na(da.wide)] <- 0
#create long data with null observations
da.long <- gather(da.wide, species, abundance, argo:yero)
#remove replicates and duplicates from ysi
ysi <- ysi[!(ysi$replicate == 2), ]
ysi$replicate <- NULL
ysi <- unique(ysi[ ,1:10])
#create temperature wide data from ysi data; can do with salinity and ph also
temp.wide <- ysi
#remove other variables(ph, salinity)
temp.wide[8:10] <- NULL
#remove depth
temp.wide[6] <- NULL
#spread to wide format
temp.wide %<>%
group_by(site, year, month, day, location) %>%
spread("location", "temp")
#merge temperature data
da.wide.habitat.env <- merge(da.wide, temp.wide,
by = c("year", "month", "day", "site"),
all = TRUE)
#merge habitat data
da.wide.habitat.env <- merge(da.wide.habitat.env, habitat, by = "site")
da.wide.habitat.env[5] <- NULL
#assign levels to month
da.wide.habitat.env$month <- factor(da.wide.habitat.env$month,
levels = c(1,2,3,4,5,6,7,8,9,10,11,12),
labels = c("Jan",
"Feb",
"Mar",
"Apr",
"May",
"Jun",
"Jul",
"Aug",
"Sep",
"Oct",
"Nov*",
"Dec"))
da.long$month <- factor(da.long$month,
levels = c(1,2,3,4,5,6,7,8,9,10,11,12),
labels = c("Jan",
"Feb",
"Mar",
"Apr",
"May",
"Jun",
"Jul",
"Aug",
"Sep",
"Oct",
"Nov*",
"Dec"))
#create date variable for TropFishR
#multiple monthly sample events justify bimonthly sample units
dl.month.dup <- dl
dl.month.dup$bimonth <- if_else(dl.month.dup$day < 15, 7, 22)
dl.month.dup$date <- ymd(paste(dl.month.dup$year, dl.month.dup$month, dl.month.dup$bimonth))
#assign levels to month
dl$month <- factor(dl$month,
levels = c(1,2,3,4,5,6,7,8,9,10,11,12),
labels = c("Jan","Feb","Mar","Apr","May","Jun","Jul","Aug","Sep","Oct","Nov","Dec"))
#average lengths
dl.year.month.core <- dl
dl.year.month.core <-
filter(dl.year.month.core, site == "ssp" | site == "hdo" | site == "pba" | site == "chp" | site == "ppo" | site == "wfb") %>%
group_by(year, month, species) %>%
summarise(se = sd(length)/sqrt((length(month))), length = mean(length))Individual Species
The following presents seasonal abundance and length data for several of the more common species. Abundance for individual species is calculated as the mean of total catch for a site visit(seine sets are lumped together). Length is fork length(FL). Error bars represent standard error of the mean. Summaries are constrained to core sites only.
Length frequency is presented both as absolute (all years and all months) and seasonally (bimonthly). Seasonal mean length is also presented where applicable (only one cohort present).
#filter core sites; generate SE
da.long.year.month.core <- da.long %>%
filter(site == "ssp" | site == "hdo" | site == "pba" | site == "chp" | site == "ppo" | site == "wfb") %>%
group_by(year, month, species) %>%
summarise(se = sd(abundance)/sqrt((length(species))), abundance = mean(abundance))
#length frequency graphics function by year month for individual species
##function of Species and Length filter(outliers)
lfq_year_month <- function(Species, Length) {
plotlfq <- dl.temp <- filter(dl.month.dup, species == Species)
dl.temp <- filter(dl.temp, site == "ssp" | site == "hdo" | site == "pba" | site == "chp" | site == "ppo" | site == "wfb")
dl.temp <- filter(dl.temp, length < Length)
dl.lfq <- lfqCreate(dl.temp, Lname = "length", Dname = "date", Fname = NA, bin_size = 2,
length_unit = "cm", plus_group = FALSE, aggregate_dates = FALSE,
plot = FALSE)
plot(dl.lfq, Fname = "catch", ylab = "Length (mm)")
}
#seasonal mean length function for any species with standard error
#only useful for species with single cohorts; doesn't work for pipefish, perch, herring, etc
mean_length_seasonal <- function(Species) {
filter(dl.year.month.core, species == Species) %>%
ggplot() +
aes(x = month, y = length, colour = factor(year), group = year) +
geom_point(size = 2.5, position = position_dodge(width = 0.25)) +
geom_line(position = position_dodge(width = 0.25)) +
geom_errorbar(aes(ymin = length - se,
ymax = length + se),
width = 0,
size = 1,
position = position_dodge(width = 0.25)) +
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
#legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1)) +
labs(x = "Month", y = expression("Mean Length"), colour = "year")
}
#the following is for less frequent species(indicator species)
abundance_seasonal <- function(Species){
filter(da.long.year.month.core, species == Species) %>%
filter(year != "2018") %>%
filter(abundance > 0) %>%
ggplot() +
aes(x = month, y = abundance, colour = species, group = species) +
geom_point(size = 2.5, position = position_dodge(width = 0.25)) +
geom_line( position = position_dodge(width = 0.25)) +
geom_errorbar(aes(ymin = abundance - se,
ymax = abundance + se),
width = 0,
size = 1,
position = position_dodge(width = 0.25))+
facet_wrap(~year)+
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1)) +
labs(x = "Month", y = expression(Abundance ~ "(mean count/site visit)"))
}Rockfish
All rockfish (abundance)
- core sites
- error bars are standard error
filter(da.long.year.month.core, species == "byro" | species == "cqbr" | species == "unro" | species == "coro" | species == "blro"
| species == "caro" | species == "boro" | species == "qbro" | species == "vero") %>%
filter(year != "2018") %>%
filter(abundance > 0) %>%
ggplot() +
aes(x = month, y = abundance, colour = factor(species, labels = c("Black",
"Black/Yellow",
"Boccacio",
"Canary",
"Copper",
"Copper/Quillback",
"Quillback",
"Unknown",
"Vermillion")), group = species) +
geom_point(size = 2.5, position = position_dodge(width = 0.25)) +
geom_line( position = position_dodge(width = 0.25)) +
geom_errorbar(aes(ymin = abundance - se,
ymax = abundance + se),
width = 0,
size = 1,
position = position_dodge(width = 0.25))+
facet_wrap(~year)+
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
#legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1)) +
labs(x = "Month", y = expression("Rockfish Abundance" ~ "(count/site visit)"), colour = "Species")
Excluding Black/Yellowtail Complex (abundance)
- core sites
- error bars are standard error
The most abundant species are removed to graphically inflate species of lower abundance.
filter(da.long.year.month.core, species == "cqbr" | species == "unro" | species == "coro" | species == "blro"
| species == "caro" | species == "boro" | species == "qbro" | species == "vero") %>%
filter(year != "2018") %>%
filter(abundance > 0) %>%
ggplot() +
aes(x = month, y = abundance, colour = factor(species, labels = c("Black",
"Boccacio",
"Canary",
"Copper",
"Copper/Quillback",
"Quillback",
"Unknown",
"Vermillion")), group = species) +
geom_point(size = 2.5, position = position_dodge(width = 0.25)) +
geom_line( position = position_dodge(width = 0.25)) +
geom_errorbar(aes(ymin = abundance - se,
ymax = abundance + se),
width = 0,
size = 1,
position = position_dodge(width = 0.25))+
facet_wrap(~year)+
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
#legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1)) +
labs(x = "Month", y = expression("Rockfish Abundance" ~ "(count/site visit)"), colour = "Species")
Excluding Black/Yellowtail Complex, Copper/Quillback (abundance)
- core sites
- error bars are standard error
The most abundant species are removed to graphically inflate species of lower abundance.
#rockfish seasonal; byro, cqbr, yellowtail excluded
filter(da.long.year.month.core, species == "unro" | species == "coro" | species == "blro"
| species == "caro" | species == "boro" | species == "qbro") %>%
filter(year != "2018") %>%
filter(abundance > 0) %>%
ggplot() +
aes(x = month, y = abundance, colour = factor(species, labels = c("Black",
"Boccacio",
"Canary",
"Quillback",
"Unknown",
"Vermillion")), group = species) +
geom_point(size = 2.5, position = position_dodge(width = 0.25)) +
geom_line( position = position_dodge(width = 0.25)) +
geom_errorbar(aes(ymin = abundance - se,
ymax = abundance + se),
width = 0,
size = 1,
position = position_dodge(width = 0.25))+
facet_wrap(~year)+
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
#legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1)) +
labs(x = "Month", y = expression("Rockfish Abundance" ~ "(count/site visit)"), colour = "Species")
Copper-quillback Complex (length frequency)
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
#length frequency graphics function by year month for individual species
##function of Species and Length filter(outliers)
lfq_year_month("cqbr", 1000)
Seasonal Mean Length
- core sites
- error bars are standard error
#execute above function
#only useful for species with single cohorts; doesn't work for pipefish, perch, herring, etc
mean_length_seasonal("cqbr")
Black-Yellow Complex (length frequency)
Length frequency
- core sites
- absolute (all years combined) and bimonthly
#length frequency graphics function by year month for individual species
##function of Species and Length filter(outliers)
lfq_year_month("byro", 1000)
Seasonal Mean Length
- core sites
- error bars are standard error
#execute above function
#only useful for species with single cohorts; doesn't work for pipefish, perch, herring, etc
mean_length_seasonal("byro")
Other Species (no apparent seasonal differences)
These species show no apparent seasonality with respects to numbers or recruitment. There appears to be a similar number from year to year and a similar size distribution.
Flatfish
Seasonal Abundance
- core sites
- error bars are standard error
#from species_individual.R
filter(da.long.year.month.core, species == "enso" | species == "saso" |species == "pasa") %>%
filter(year != "2018") %>%
filter(abundance > 0) %>%
ggplot() +
aes(x = month, y = abundance, colour = factor(species, labels = c("English Sole",
"Sanddab",
"Sand Sole")), group = species) +
geom_point(size = 2.5, position = position_dodge(width = 0.25)) +
geom_line( position = position_dodge(width = 0.25)) +
geom_errorbar(aes(ymin = abundance - se,
ymax = abundance + se),
width = 0,
size = 1,
position = position_dodge(width = 0.25))+
facet_wrap(~year)+
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
#legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1)) +
labs(x = "Month", y = expression("Flatfish Abundance" ~ "(count/site visit)"), colour = "Species")
English sole
Length frequency
- core sites
- absolute (all years combined) and bimonthly
#length frequency graphics function by year month for individual species
##function of Species and Length filter(outliers)
lfq_year_month("enso", 200)
Seasonal Mean Length
- core sites
- error bars are standard error
#execute above function
#only useful for species with single cohorts; doesn't work for pipefish, perch, herring, etc
mean_length_seasonal("enso")
Sand sole
Length frequency
- core sites
- absolute (all years combined) and bimonthly
#length frequency graphics function by year month for individual species
##function of Species and Length filter(outliers)
lfq_year_month("saso", 200)
Seasonal Mean Length
- core sites
- error bars are standard error
#execute above function
#only useful for species with single cohorts; doesn't work for pipefish, perch, herring, etc
mean_length_seasonal("saso")
Sanddab
Length frequency
- core sites
- absolute (all years combined) and bimonthly
#length frequency graphics function by year month for individual species
##function of Species and Length filter(outliers)
lfq_year_month("pasa", 200)
Seasonal Mean Length
- core sites
- error bars are standard error
#execute above function
#only useful for species with single cohorts; doesn't work for pipefish, perch, herring, etc
mean_length_seasonal("pasa")
Sandlance
Seasonal Abundance
- core sites
- error bars are standard error
abundance_seasonal("sand")
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
lfq_year_month("sand", 1000)
Lingcod
Seasonal Abundance
- core sites
- error bars are standard error
abundance_seasonal("ling")
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
lfq_year_month("ling", 1000)
Seasonal Mean Length
- core sites
- error bars are standard error
#execute above function
#only useful for species with single cohorts; doesn't work for pipefish, perch, herring, etc
mean_length_seasonal("ling")
Herring
Seasonal Abundance
- core sites
- error bars are standard error
abundance_seasonal("herr")
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
lfq_year_month("herr", 1000)
Shiner Perch
Seasonal Abundance
- core sites
- error bars are standard error
abundance_seasonal("shpe")
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
lfq_year_month("shpe", 1000)
Bay Pipefish
Seasonal Abundance
- core sites
- error bars are standard error
abundance_seasonal("bapi")
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
lfq_year_month("bapi", 1000)
Kelp greenling
Seasonal Abundance
- core sites
- error bars are standard error
abundance_seasonal("kegr")
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
lfq_year_month("kegr", 1000)
Tidepool Sculpin
Seasonal Abundance
- core sites
- error bars are standard error
abundance_seasonal("tisc")
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
lfq_year_month("tisc", 100)
Buffalo Sculpin
Seasonal Abundance
- core sites
- error bars are standard error
abundance_seasonal("busc")
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
lfq_year_month("busc", 150)
Three-Spine Stickleback
Seasonal Abundance
- core sites
- error bars are standard error
abundance_seasonal("thst")
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
lfq_year_month("thst", 1000)
Staghorn Sculpin
Seasonal Abundance
- core sites
- error bars are standard error
abundance_seasonal("stsc")
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
lfq_year_month("stsc", 1000)
Black-Eyed Goby
Seasonal Abundance
- core sites
- error bars are standard error
abundance_seasonal("blgo")
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
lfq_year_month("blgo", 1000)
Bay Goby
Seasonal Abundance
- core sites
- error bars are standard error
abundance_seasonal("bago")
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
lfq_year_month("bago", 1000)
Other Species (Annual differences)
These species show strong annual diffences in abundance and/or recruitment. These could make for potential indicator species. More research into life history traits would be required.
Crescent Gunnel
Seasonal Abundance
- core sites
- error bars are standard error
abundance_seasonal("crgu")
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
lfq_year_month("crgu", 1000)
Penpoint Gunnel
Seasonal Abundance
- core sites
- error bars are standard error
abundance_seasonal("pegu")
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
lfq_year_month("pegu", 1000)
Great Sculpin
Seasonal Abundance
- core sites
- error bars are standard error
abundance_seasonal("grsc")
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
lfq_year_month("grsc", 1000)
Silverspotted Sculpin
Seasonal Abundance
- core sites
- error bars are standard error
abundance_seasonal("sisc")
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
lfq_year_month("sisc", 1000)
Snake Prickleback
Seasonal Abundance
- core sites
- error bars are standard error
abundance_seasonal("snpr")
Length Frequency
- core sites
- absolute (all years combined) and bimonthly
lfq_year_month("snpr", 1000)
Species Diversity
#create derived variables for Shannon index and species richness
#new column; shannon H
da.wide.habitat.env$shan <- diversity(
da.wide.habitat.env[(which(colnames(da.wide.habitat.env) == "argo")) :
(which(colnames(da.wide.habitat.env) == "yero"))],
index = "shannon")
#new column; specnumber
da.wide.habitat.env$abundance<- specnumber(
da.wide.habitat.env[(which(colnames(da.wide.habitat.env) == "argo")) :
(which(colnames(da.wide.habitat.env) == "yero"))])Species Accumulation Curve
(Kindt’s exact method)
This demonstrates that we have pretty much plateaued in species accumulation. Although, we occasionally find a new species, we only found two new species this year.
specaccum(da.wide.habitat.env[
(which(colnames(da.wide.habitat.env) == "argo")) :
(which(colnames(da.wide.habitat.env) == "yero"))]) %>%
plot(ci.type="polygon", ci.col="lightyellow", xlab = "site visits", ylab = "species number")
Species Diversity (temporal and spatial)
The following figures are two metrics (species richness and Shannons index) of diversity filtered, sorted amd displayed in several ways. Richness is calculated as the total number of species observed. Shannonds H take into account both richness and evenness.
Nov*: This sampling event is unique in that it was done late at night and the species comosition is somewhat distinct. Results should be interpreted with caution.
All sites, all years
#diversity(spp.richness and shannons) by site and month; all sites
da.month.diversity<- (group_by(da.wide.habitat.env, month, site))%>%
summarise_at(vars(shan:abundance), mean)Species Richness
- all sites; all years
- night sampling(november)
#spp.richness
ggplot(da.month.diversity) +
aes(x = month, y = abundance, colour = site, group = site) +
geom_point() +
geom_line() +
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1))+
labs(x = "Month", y = expression("Richness" ~ "(species count/site visit)"))
Shannons Index
all sites; all years
#shannons
ggplot(da.month.diversity) +
aes(x = month, y = shan, colour = site, group = site) +
geom_point() +
geom_line() +
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1)) +
labs(x = "Month", y = expression("Shannon's H" ~ "(Shannons H/site visit)"))
Core sites, all years
#diversity(spp.richness and shannons) by site and month; core sites
da.month.diversity.core<- (group_by(da.wide.habitat.env, month, site))%>%
summarise_at(vars(shan:abundance), mean) %>%
filter(site == "ssp" | site == "hdo" | site == "pba" | site == "chp" | site == "ppo" | site == "wfb")Species Richness
core sites; all years
#spp.richness
ggplot(da.month.diversity.core) +
aes(x = month, y = abundance, colour = site, group = site) +
geom_point() +
geom_line() +
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
#legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1))+
labs(x = "Month", y = expression("Richness" ~ "(species count/site visit)"))
Shannons Index
core sites; all years
#shannons
ggplot(da.month.diversity.core) +
aes(x = month, y = shan, colour = site, group = site) +
geom_point() +
geom_line() +
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
#legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1)) +
labs(x = "Month", y = expression("Shannon's H" ~ "(Shannons H/site visit)"))
All sites by year
#diversity(spp.richness and shannons) by year, site and month; all sites
da.year<- (group_by(da.wide.habitat.env, year, month, site))%>%
summarise_at(vars(shan:abundance), mean)Species Richness
all sites; all years
#spp.richness
ggplot(da.year) +
aes(x = month, y = abundance, colour = site, group = site) +
geom_point() +
geom_line() +
facet_wrap(~year) +
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1)) +
labs(x = "Month", y = expression("Richness" ~ "(species count/site visit)"))
Shannons Index
all sites; all years
#shannons H
ggplot(da.year) +
aes(x = month, y = shan, colour = site, group = site) +
geom_point() +
geom_line() +
facet_wrap(~year) +
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1)) +
labs(x = "Month", y = expression("Shannon's H" ~ "(Shannons H/site visit)"))
Core sites by year
#diversity(spp.richness and shannons) by year, site and month; core sites
da.year<- (group_by(da.wide.habitat.env, year, month, site))%>%
summarise_at(vars(shan:abundance), mean) %>%
filter(site == "ssp" | site == "hdo" | site == "pba" | site == "chp" | site == "ppo" | site == "wfb")Species Richness
core sites; all years
#spp.richness
ggplot(da.year) +
aes(x = month, y = abundance, colour = site, group = site) +
geom_point() +
geom_line() +
facet_wrap(~year) +
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
#legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1)) +
labs(x = "Month", y = expression("Richness" ~ "(species count/site visit)"))
Shannons Index
core sites; all years
#shannons H
ggplot(da.year) +
aes(x = month, y = shan, colour = site, group = site) +
geom_point() +
geom_line() +
facet_wrap(~year) +
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
#legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1)) +
labs(x = "Month", y = expression("Shannon's H" ~ "(Shannons H/site visit)"))
Species Diversity and Total Abundance (habitat delineation)
The following are several representations of diverity and abundance by basic habitat type. Habitat was categorized by coarse metrics of substrate and vegetation. Ex) “unal.gravel” = unidentified algae with primarily gravel substrate. Substrate appears to be a good proxy for exposure. Exposed sites have a coarser substrate (sand and gravel), where sheltered sites have a mud (silt and clay) substrate.
Only June, July and August data were used as richness appears to be stable and at a plateau. Richness is calculated as the total number of species observed. Shannonds H take into account both richness and evenness.
Total fish counts
summer months (June, July and August)
#diversity by habitat type
#total fish abundance
da.long.habitat <- merge(da.long, habitat, by = "site", incomparables = NA)
group_by(da.long.habitat, year, month, site, veg_subst) %>%
summarise(abundance1 = sum(abundance, na.rm = TRUE)) %>%
group_by(year, month, veg_subst) %>%
summarise(abundance = mean(abundance1), sd(abundance1)) %>%
filter(month == "Aug" | month == "Jul" | month == "Jun") %>%
ggplot() +
aes(x = veg_subst, y = abundance) +
geom_boxplot() +
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1))+
labs(x = "vegetation and substrate", y = expression("Fish abundance" ~ "(total fish count/site visit)"))
Shannon H
summer months (June, July and August)
# shannon H by habitat
filter(da.wide.habitat.env, month == "Aug" | month == "Jul" | month == "Jun") %>%
ggplot() +
aes(x = veg_subst, y = shan) +
geom_boxplot() +
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1))+
labs(x = "vegetation and substrate", y = expression("Shannons H"))
Species richness
summer months (June, July and August)
# richness by habitat
filter(da.wide.habitat.env, month == "Aug" | month == "Jul" | month == "Jun") %>%
ggplot() +
aes(x = veg_subst, y = abundance) +
geom_boxplot() +
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1))+
labs(x = "vegetation and substrate", y = expression("species count"))
Community Dissimilarity
NMDS (Bray-Curtis) ordinations of community composition for the community data.
#extract species observations and sites only
#summarise by site/month (analyses cannot account for day)
#merge habitat data
da.obs <- da.wide.habitat.env
da.obs <- da.obs[1:(which(colnames(da.obs) == "yero"))]
da.obs%<>%
group_by(site, year, month)
da.obs <- summarise_all(da.obs, funs(mean))
da.obs <- merge(da.obs, habitat, by = "site")Between Habitats
NMDS (Bray-Curtis) ordinations of community composition by year in the month of July. Grouped by habitat type. Again, habitat was categorized by coarse metrics of substrate and vegetation. Ex) “unal.gravel” = unidentified algae with primarily gravel substrate. Substrate appears to be a good proxy for exposure. Exposed sites have a coarser substrate (sand and gravel), where sheltered sites have a mud (silt and clay) substrate.
2014
all sites; July
#2014 July NMDS
da.wide.habitat.env.july2014 <- filter(da.obs,
year == "2014",
month == "Jul")
da.wide.habitat.env.july2014 <- column_to_rownames(da.wide.habitat.env.july2014,
var = "site")
nmds.july2014 <- metaMDS(da.wide.habitat.env.july2014[(which(colnames(da.wide.habitat.env.july2014) == "argo")) :
(which(colnames(da.wide.habitat.env.july2014) == "yero"))],
k=3)
ordiplot(nmds.july2014,
display="site",
type="n",
xlim = c(-2, 2))
points(nmds.july2014,
col="black",
pch = (as.integer(da.wide.habitat.env.july2014$veg_subst)))
ordihull(nmds.july2014, da.wide.habitat.env.july2014$veg_subst,
scaling = "symmetric",
lty=(as.integer(da.wide.habitat.env.july2014$veg_subst)))
legend("topright", levels(da.wide.habitat.env.july2014$veg_subst),
pch=1:(length(levels(da.wide.habitat.env.july2014$veg_subst))),
lty=1:(length(levels(da.wide.habitat.env.july2014$veg_subst))))
2015
all sites; July
#2015 July NMDS
da.wide.habitat.env.july2015 <- filter(da.obs,
year == "2015",
month == "Jul")
da.wide.habitat.env.july2015 <- column_to_rownames(da.wide.habitat.env.july2015,
var = "site")
nmds.july2015 <- metaMDS(da.wide.habitat.env.july2015[(which(colnames(da.wide.habitat.env.july2015) == "argo")) :
(which(colnames(da.wide.habitat.env.july2015) == "yero"))],
k=3)
ordiplot(nmds.july2015,
display="site",
type="n",
xlim = c(-2, 2))
points(nmds.july2015,
col="black",
pch = (as.integer(da.wide.habitat.env.july2015$veg_subst)))
ordihull(nmds.july2015, da.wide.habitat.env.july2015$veg_subst,
scaling = "symmetric",
lty=(as.integer(da.wide.habitat.env.july2015$veg_subst)))
legend("topright", levels(da.wide.habitat.env.july2015$veg_subst),
pch=1:(length(levels(da.wide.habitat.env.july2015$veg_subst))),
lty=1:(length(levels(da.wide.habitat.env.july2015$veg_subst))))
2016
all sites; July
#2016 July NMDS
da.wide.habitat.env.july2016 <- filter(da.obs,
year == "2016",
month == "Jul")
da.wide.habitat.env.july2016 <- column_to_rownames(da.wide.habitat.env.july2016,
var = "site")
nmds.july2016 <- metaMDS(da.wide.habitat.env.july2016[(which(colnames(da.wide.habitat.env.july2016) == "argo")) :
(which(colnames(da.wide.habitat.env.july2016) == "yero"))],
k=3)
ordiplot(nmds.july2016,
display="site",
type="n",
xlim = c(-2, 2))
points(nmds.july2016,
col="black",
pch = (as.integer(da.wide.habitat.env.july2016$veg_subst)))
ordihull(nmds.july2016, da.wide.habitat.env.july2016$veg_subst,
scaling = "symmetric",
lty=(as.integer(da.wide.habitat.env.july2016$veg_subst)))
legend("topright", levels(da.wide.habitat.env.july2016$veg_subst),
pch=1:(length(levels(da.wide.habitat.env.july2016$veg_subst))),
lty=1:(length(levels(da.wide.habitat.env.july2016$veg_subst))))
2017
all sites; July
#2017 July NMDS
da.wide.habitat.env.july2017 <- filter(da.obs,
year == "2017",
month == "Jul")
da.wide.habitat.env.july2017 <- column_to_rownames(da.wide.habitat.env.july2017,
var = "site")
nmds.july2017 <- metaMDS(da.wide.habitat.env.july2017[(which(colnames(da.wide.habitat.env.july2017) == "argo")) :
(which(colnames(da.wide.habitat.env.july2017) == "yero"))],
k=3)
ordiplot(nmds.july2017,
display="sites",
type="n",
xlim = c(-2, 2))
points(nmds.july2017,
col="black",
pch = (as.integer(da.wide.habitat.env.july2017$veg_subst)))
ordihull(nmds.july2017, da.wide.habitat.env.july2017$veg_subst,
scaling = "symmetric",
lty=(as.integer(da.wide.habitat.env.july2017$veg_subst)))
legend("topright", levels(da.wide.habitat.env.july2017$veg_subst),
pch=1:(length(levels(da.wide.habitat.env.july2017$veg_subst))),
lty=1:(length(levels(da.wide.habitat.env.july2017$veg_subst)))) 
Between Years and Habitat
NMDS (Bray-Curtis) ordinations of community composition in summer months (June, July and August). Grouped by year and habitat type. Only annual sites were used. Annual sites are those that were visited every year and includes core sites.
All habitats
annual sites; summer months
#NMDS by year;summer months(June, July, August); all habitats
da.wide.habitat.env.trial <- filter(da.obs,
month == "Jun" | month == "Jul" | month == "Aug",
site == "chp" | site == "fan1" | site == "fan3" | site == "gog1" | site == "hdp" |
site == "kis2" | site == "pba" | site == "ppo" | site == "sni1" | site == "sni2" |
site == "ssp" | site == "wfb")
da.wide.habitat.env.trial$year <- as.factor(da.wide.habitat.env.trial$year)
nmds.trial <- metaMDS(da.wide.habitat.env.trial[(which(colnames(da.wide.habitat.env.trial) == "argo")) :
(which(colnames(da.wide.habitat.env.trial) == "yero"))],
k=3)
ordiplot(nmds.trial,
display="site",
type="n")
points(nmds.trial,
col="black",
pch = (as.integer(da.wide.habitat.env.trial$year)))
ordihull(nmds.trial, da.wide.habitat.env.trial$year,
scaling = "symmetric",
lty=(as.integer(da.wide.habitat.env.trial$year)))
legend("topright", levels(da.wide.habitat.env.trial$year),
pch=1:(length(levels(da.wide.habitat.env.trial$year))),
lty=1:(length(levels(da.wide.habitat.env.trial$year))))
All seagrass
annual sites; summer months
##NMDS by year; summer months(June, July, August); all seagrass
da.wide.habitat.env.trial <- filter(da.obs,
month == "Jun" | month == "Jul" | month == "Aug",
site == "chp" | site == "fan1" | site == "fan3" | site == "gog1" | site == "hdp" |
site == "kis2" | site == "pba" | site == "ppo" | site == "sni1" | site == "sni2" |
site == "ssp" | site == "wfb",
veg_subst == "zostera.mud" | veg_subst == "zostera.sand" | veg_subst == "zostera,gravel")
da.wide.habitat.env.trial$year <- as.factor(da.wide.habitat.env.trial$year)
nmds.trial <- metaMDS(da.wide.habitat.env.trial[(which(colnames(da.wide.habitat.env.trial) == "argo")) :
(which(colnames(da.wide.habitat.env.trial) == "yero"))],
k=3)
ordiplot(nmds.trial,
display="site",
type="n")
points(nmds.trial,
col="black",
pch = (as.integer(da.wide.habitat.env.trial$year)))
ordihull(nmds.trial, da.wide.habitat.env.trial$year,
scaling = "symmetric",
lty=(as.integer(da.wide.habitat.env.trial$year)))
legend("topright", levels(da.wide.habitat.env.trial$year),
pch=1:(length(levels(da.wide.habitat.env.trial$year))),
lty=1:(length(levels(da.wide.habitat.env.trial$year))))
Seagrass Sheltered
annual sites; summer months
##NMDS by year; summer months(June, July, August); sheltered seagrass
da.wide.habitat.env.trial <- filter(da.obs,
month == "Jun" | month == "Jul" | month == "Aug",
site == "chp" | site == "fan1" | site == "fan3" | site == "gog1" | site == "hdp" |
site == "kis2" | site == "pba" | site == "ppo" | site == "sni1" | site == "sni2" |
site == "ssp" | site == "wfb",
veg_subst == "zostera.mud")
da.wide.habitat.env.trial$year <- as.factor(da.wide.habitat.env.trial$year)
nmds.trial <- metaMDS(da.wide.habitat.env.trial[(which(colnames(da.wide.habitat.env.trial) == "argo")) :
(which(colnames(da.wide.habitat.env.trial) == "yero"))],
k=3)
ordiplot(nmds.trial,
display="site",
type="n")
points(nmds.trial,
col="black",
pch = (as.integer(da.wide.habitat.env.trial$year)))
ordihull(nmds.trial, da.wide.habitat.env.trial$year,
scaling = "symmetric",
lty=(as.integer(da.wide.habitat.env.trial$year)))
legend("topright", levels(da.wide.habitat.env.trial$year),
pch=1:(length(levels(da.wide.habitat.env.trial$year))),
lty=1:(length(levels(da.wide.habitat.env.trial$year))))
Seagrass Exposed
annual sites; summer months
##NMDS by year; summer months(June, July, August); exposed seagrass
da.wide.habitat.env.trial <- filter(da.obs,
month == "Jun" | month == "Jul" | month == "Aug",
site == "chp" | site == "fan1" | site == "fan3" | site == "gog1" | site == "hdp" |
site == "kis2" | site == "pba" | site == "ppo" | site == "sni1" | site == "sni2" |
site == "ssp" | site == "wfb",
veg_subst == "zostera.sand" | veg_subst == "zostera.gravel")
da.wide.habitat.env.trial$year <- as.factor(da.wide.habitat.env.trial$year)
nmds.trial <- metaMDS(da.wide.habitat.env.trial[(which(colnames(da.wide.habitat.env.trial) == "argo")) :
(which(colnames(da.wide.habitat.env.trial) == "yero"))],
k=3)
ordiplot(nmds.trial,
display="site",
type="n")
points(nmds.trial,
col="black",
pch = (as.integer(da.wide.habitat.env.trial$year)))
ordihull(nmds.trial, da.wide.habitat.env.trial$year,
scaling = "symmetric",
lty=(as.integer(da.wide.habitat.env.trial$year)))
legend("topright", levels(da.wide.habitat.env.trial$year),
pch=1:(length(levels(da.wide.habitat.env.trial$year))),
lty=1:(length(levels(da.wide.habitat.env.trial$year))))
Sand
annual sites; summer months
##NMDS by year; summer months(June, July, August); sandy habitat
da.wide.habitat.env.trial <- filter(da.obs,
month == "Jun" | month == "Jul" | month == "Aug",
site == "chp" | site == "fan1" | site == "fan3" | site == "gog1" | site == "hdp" |
site == "kis2" | site == "pba" | site == "ppo" | site == "sni1" | site == "sni2" |
site == "ssp" | site == "wfb",
veg_subst == "sand")
da.wide.habitat.env.trial$year <- as.factor(da.wide.habitat.env.trial$year)
nmds.trial <- metaMDS(da.wide.habitat.env.trial[(which(colnames(da.wide.habitat.env.trial) == "argo")) :
(which(colnames(da.wide.habitat.env.trial) == "yero"))],
k=3)
ordiplot(nmds.trial,
display="site",
type="n")
points(nmds.trial,
col="black",
pch = (as.integer(da.wide.habitat.env.trial$year)))
ordihull(nmds.trial, da.wide.habitat.env.trial$year,
scaling = "symmetric",
lty=(as.integer(da.wide.habitat.env.trial$year)))
legend("topright", levels(da.wide.habitat.env.trial$year),
pch=1:(length(levels(da.wide.habitat.env.trial$year))),
lty=1:(length(levels(da.wide.habitat.env.trial$year))))
Temperature
Several summaries of temperature (1 meter depth) filtered, sorted and displayed in several ways.
#create temperature variable
temp.summary<-
group_by(da.wide.habitat.env, year, month, site, veg_subst) %>%
summarise(temp = mean(b2))Monthly Temperature
error bars are standard error
#monthly temperature by year and habitat type
group_by(temp.summary, year, month, veg_subst) %>%
summarise(se = sd(temp)/sqrt((length(veg_subst))), temp = mean(temp), n()) %>%
filter( year == "2015" | year == "2016" | year == "2017") %>%
ggplot() +
aes(x = month, y = temp, colour = factor(veg_subst, labels = c("macro.sand",
"mud",
"sand",
"ulva.sand",
"unal.cobble",
"unal.gravel",
"unal.mud",
"unal.sand",
"zostera.gravel",
"zostera.mud",
"zostera.sand")), group = veg_subst) +
geom_point(size = 2.5, position = position_dodge(width = 0.25)) +
geom_line( position = position_dodge(width = 0.25)) +
geom_errorbar(aes(ymin = temp - se,
ymax = temp + se),
width = 0,
size = 1,
position = position_dodge(width = 0.25))+
facet_wrap(~year)+
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
#legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1)) +
labs(x = "Month", y = expression("temperature"), colour = "veg_subst")
Temperature by Habitat
summer; all sites
#Temperature by habitat type
group_by(temp.summary, year, month, veg_subst)%>%
summarise(se = sd(temp)/sqrt((length(veg_subst))), temp = mean(temp), n()) %>%
filter(month == "Aug" | month == "Jul" | month == "Jun") %>%
ggplot() +
aes(x = veg_subst, y = temp) +
geom_boxplot() +
facet_wrap(~year) +
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
legend.position = "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1))+
labs(x = "vegetation and substrate", y = expression("temperature"))
Temperature by Habitat
Summer; core sites
error bars are standard error
#temperature by habitat type
filter(temp.summary,
year != "2014",
month == "Jun" | month == "Jul",
site == "ssp" | site == "hdo" | site == "pba" | site == "chp" | site == "ppo" | site == "wfb") %>%
group_by(year, veg_subst)%>%
summarise(se = sd(temp, na.rm = TRUE)/sqrt((length(veg_subst))),
temp = mean(temp, na.rm = TRUE), n()) %>%
ggplot() +
aes(x = veg_subst, y = temp, colour = factor(year, labels = c("2015",
"2016",
"2017")), group = year) +
geom_point(size = 2.5,
position = position_dodge(width = 0.25)) +
geom_errorbar(aes(ymin = temp - se,
ymax = temp + se),
width = 0,
size = 1,
position = position_dodge(width = 0.25)) +
theme_bw() +
theme(axis.text.x = element_text(size = 10, angle = 45, hjust = 1),
#legend.position= "none",
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
text = element_text(size = 13, vjust = 0.1))+
labs(x = "vegetation and substrate", y = expression("temperature"), color = "Year")